This invention relates to active matrix electro-optic display devices, and more particularly relates to such devices having a capacitive element associated with each pixel of the matrix, and also relates to color projection apparatus incorporating these devices.
In one form of color projection television (PTV) in widespread use today, a full color display is formed by superimposing three different (red, blue and green) monochrome images on a projection screen. While these monochrome images are usually formed using cathode ray tubes (CRTs), it has also been proposed to use active matrix electro-optic displays such as thin film transistor (TFT) switched twisted nematic liquid crystal displays (TN LCD) in place of the CRTs to generate these images.
In a LCD cell operating in the transmission mode (sometimes referred to as a "light valve"), a liquid crystal (LC) material is sandwiched between two optically transparent walls, the inner surfaces of which each carry an orienting layer in contact with the LC, for promoting a preferential alignment of the LC molecules adjacent to the layers. Because the LC molecules also tend to align with one another, as well as with the orienting layers, a twist can be imparted to the LC by assembling the cell with a non-parallel orientation of these layers. For example, arranging the layers with their alignment directions orthogonal to one another produces a 90.degree. twist of a nematic LC from one surface to the other. This twist enables the LC to rotate linearly polarized light by 90.degree., that the amount of light transmitted by the device can be controlled by an exit polarizer, called an analyzer. Furthermore, the LC can be untwisted by application of a voltage, so that, for example, light blocked by an analyzer having its polarization direction oriented parallel to that of the entrance polarizer, can be passed by application of an appropriate voltage to the LC cell.
A two-dimensional array of such cells, each cell individually addressable through a matrix of row and column electrodes, can be used to build up a display such as a video image, where each cell constitutes a pixel of information. During the repetitive "scanning" (sequential addressing) of the pixel rows, e.g., in response to a video signal, it has been found advantageous to reduce cross-talk between the pixels by providing a separate switch for each pixel. This so-called "active matrix" addressing scheme depends for its success on the ability of the pixels to hold their charge received during addressing, by virtue of their inherent capacitances, until they are readdressed.
However, charge can leak off the pixels, eg., through the liquid crystal material or through the switch. It is thus common in the design of the pixel to add extra capacitance in order to improve its charge storage capacity. One way of introducing the extra capacitance that has been used is to extend the transparent pixel electrode to overlap the adjacent row electrode above the cross-over dielectric between the row and column electrodes. Since the row electrode remains at a fixed potential except when being addressed, it becomes a convenient base plate for the extra capacitor.
FIG. 1 shows such a prior art arrangement for a pixel of an active matrix TFT switched LCD light valve, in which an array of thin transparent pixel electrodes 10, here of indium tin oxide (ITO), are arranged in a matrix of row and column electrodes, 12 and 14 respectively, separated by a cross-over dielectric 16, here of chemically vapor deposited (CVD) oxide. Each pixel 10 is accessed through a switch such as a TFT 18 (only one of which is shown in the Figure). The switch includes a polysilicon layer 20, one or more gate electrodes 22 and 24, which overlie channel regions (not shown) in layer 20 and which are connected to the polysilicon row electrode (known as a gate line) 12, a gate dielectric under the gate electrodes (not shown), and source and drain regions 26 and 28, which are respectively connected to the aluminum column electrode (known as a source line) 14 and to the pixel electrode 10.
The extra capacitor 29 is formed by extending the ITO 10 over the adjacent gate line. The cross-over dielectric 16 provides a convenient dielectric for the extra capacitor. However, the CVD layer 16 is usually relatively thick (about 1 micron). Because of this, the amount of extra capacitance that can be obtained in this manner is about the same as the capacitance of the liquid crystal pixel. In projection television, for example, this may not be adequate since the pixels are typically very small (for example 250,000 pixels may be arranged in a 2 square inch area), and hence so is the LC capacitance of each pixel.
Seiko-Epson has approached this problem by forming a thin polysilicon region 30 (see FIG. 2) at the same time that the channel poly silicon 20 is formed for the TFT, and oxidizing it at the same time that the gate dielectric is formed for the TFT 18. Another thick polysilicon line 32 is then formed at the same time that the gate line 12 is formed, extending over the thin polysilicon region 30 to form the extra capacitor 40. The dielectric 34 for the extra capacitor is a relatively thin thermally oxidized layer about 1,000 angstroms thick. The capacitance thus provided is about 5 to 7 times the pixel capacitance. The extra line 32 extends across the entire array in order to allow a bias voltage to be applied to the capacitor so that the undoped thin polysilicon region underlying the extra line is charged to make it conducting under all operating conditions of the TFT. A severe penalty which results from this design is that the extra line reduces the geometric transparency of the pixel by as much as 25% or more.